I hate to say it, but I have had some data loss using SEW on Dropbox. Perhaps it wouldn’t be the same with Drive (I use both, so I should test it sometime).
I think what happened was that there was sufficient time delay between some fragment of my document getting onto the cloud for SEW to perform another of its unpredictable (by the user) write operations. I may have been rapidly flipping back and forth between several scenes, too, and I type very fast at times. But I suddenly realized that some text was missing here and there, and since I’d been furiously editing all over the place there was no way to determine what might have been corrupted. I beat a hasty retreat, as they say.
I use other programs in this way all the time, on both Dropbox and Drive, and haven’t noticed anything, but that doesn’t guarantee nothing ever happened. SEW uses a semaphore system to protect a local file store (project) from being clobbered when one instance of SEW has opened it, but it’s not too hard to imagine a slow day on the interweb tubing, when a project gets closed or a connection gets severed before everything in the cloud is up to date. My fiber optic feed is usually close to 100MHz, but some intermediate segment of tubing might just happen to be an old bit of garden hose, and with all those little files being opened and closed as I edit wildly, I think a more elaborate scheme would be needed to guarantee data integrity.
So I caution anyone considering the cloud as storage for SEW projects to avoid thrashing about, and always allow plenty of time (usually a few seconds should do) before and after opening and closing scenes. And perhaps click that pale blue 3.25" disc icon whenever it’s not grey.
I wonder, sometimes, if scenarios like these are part of Google’s inspiration to save every few seconds, rather than every minute or so, but surely they must be using a much more complex scheme to protect documents being edited over potentially sluggish routes by numerous people same-ul-time-eously.